Home

Context Navigation

source: palm/trunk/SOURCE/pres.f90 @ 1683

Last change on this file since 1683 was 1683, checked in by knoop, 9 years ago
last commit documented
Property svn:keywords set to `Id`
File size: 24.3 KB

Rev	Line
[1682]	1	!> @file pres.f90
[1036]	2	!--------------------------------------------------------------------------------!
	3	! This file is part of PALM.
	4	!
	5	! PALM is free software: you can redistribute it and/or modify it under the terms
	6	! of the GNU General Public License as published by the Free Software Foundation,
	7	! either version 3 of the License, or (at your option) any later version.
	8	!
	9	! PALM is distributed in the hope that it will be useful, but WITHOUT ANY
	10	! WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
	11	! A PARTICULAR PURPOSE. See the GNU General Public License for more details.
	12	!
	13	! You should have received a copy of the GNU General Public License along with
	14	! PALM. If not, see <http://www.gnu.org/licenses/>.
	15	!
[1310]	16	! Copyright 1997-2014 Leibniz Universitaet Hannover
[1036]	17	!--------------------------------------------------------------------------------!
	18	!
[484]	19	! Current revisions:
[1212]	20	! ------------------
[1683]	21	!
	22	!
[1321]	23	! Former revisions:
	24	! -----------------
	25	! $Id: pres.f90 1683 2015-10-07 23:57:51Z knoop $
	26	!
[1683]	27	! 1682 2015-10-07 23:56:08Z knoop
	28	! Code annotations made doxygen readable
	29	!
[1576]	30	! 1575 2015-03-27 09:56:27Z raasch
	31	! poismg_fast + respective module added, adjustments for psolver-queries
	32	!
[1343]	33	! 1342 2014-03-26 17:04:47Z kanani
	34	! REAL constants defined as wp-kind
	35	!
[1321]	36	! 1320 2014-03-20 08:40:49Z raasch
[1320]	37	! ONLY-attribute added to USE-statements,
	38	! kind-parameters added to all INTEGER and REAL declaration statements,
	39	! kinds are defined in new module kinds,
	40	! old module precision_kind is removed,
	41	! revision history before 2012 removed,
	42	! comment fields (!:) to be used for variable explanations added to
	43	! all variable declaration statements
[708]	44	!
[1319]	45	! 1318 2014-03-17 13:35:16Z raasch
	46	! module interfaces removed
	47	!
[1307]	48	! 1306 2014-03-13 14:30:59Z raasch
	49	! second argument removed from routine poisfft
	50	!
[1258]	51	! 1257 2013-11-08 15:18:40Z raasch
	52	! openacc loop and loop vector clauses removed, independent clauses added,
	53	! end parallel replaced by end parallel loop
	54	!
[1222]	55	! 1221 2013-09-10 08:59:13Z raasch
	56	! openACC porting of reduction operations, loops for calculating d are
	57	! using the rflags_s_inner multiply flag instead of the nzb_s_inner loop index
	58	!
[1213]	59	! 1212 2013-08-15 08:46:27Z raasch
	60	! call of poisfft_hybrid removed
	61	!
[1118]	62	! 1117 2013-03-27 11:15:36Z suehring
	63	! Bugfix in OpenMP parallelization.
	64	!
[1114]	65	! 1113 2013-03-10 02:48:14Z raasch
	66	! GPU-porting of several loops, some loops rearranged
	67	!
[1112]	68	! 1111 2013-03-08 23:54:10Z
	69	! openACC statements added,
	70	! ibc_p_b = 2 removed
	71	!
[1093]	72	! 1092 2013-02-02 11:24:22Z raasch
	73	! unused variables removed
	74	!
[1037]	75	! 1036 2012-10-22 13:43:42Z raasch
	76	! code put under GPL (PALM 3.9)
	77	!
[1004]	78	! 1003 2012-09-14 14:35:53Z raasch
	79	! adjustment of array tend for cases with unequal subdomain sizes removed
	80	!
[1]	81	! Revision 1.1 1997/07/24 11:24:44 raasch
	82	! Initial revision
	83	!
	84	!
	85	! Description:
	86	! ------------
[1682]	87	!> Compute the divergence of the provisional velocity field. Solve the Poisson
	88	!> equation for the perturbation pressure. Compute the final velocities using
	89	!> this perturbation pressure. Compute the remaining divergence.
[1]	90	!------------------------------------------------------------------------------!
[1682]	91	SUBROUTINE pres
	92
[1]	93
[1320]	94	USE arrays_3d, &
	95	ONLY: d, ddzu, ddzu_pres, ddzw, dzw, p, p_loc, tend, u, v, w
	96
	97	USE control_parameters, &
	98	ONLY: bc_lr_cyc, bc_ns_cyc, conserve_volume_flow, dt_3d, &
	99	gathered_size, ibc_p_b, ibc_p_t, intermediate_timestep_count, &
	100	mg_switch_to_pe0_level, on_device, outflow_l, outflow_n, &
	101	outflow_r, outflow_s, psolver, simulated_time, subdomain_size, &
	102	topography, volume_flow, volume_flow_area, volume_flow_initial
	103
	104	USE cpulog, &
	105	ONLY: cpu_log, log_point, log_point_s
	106
	107	USE grid_variables, &
	108	ONLY: ddx, ddy
	109
	110	USE indices, &
	111	ONLY: nbgp, ngp_2dh_outer, nx, nxl, nxlg, nxl_mg, nxr, nxrg, nxr_mg, &
	112	ny, nys, nysg, nys_mg, nyn, nyng, nyn_mg, nzb, nzb_s_inner, &
	113	nzb_u_inner, nzb_v_inner, nzb_w_inner, nzb_2d, nzt, nzt_mg, &
	114	rflags_s_inner
	115
	116	USE kinds
	117
[1]	118	USE pegrid
	119
[1320]	120	USE poisfft_mod, &
	121	ONLY: poisfft
	122
[1575]	123	USE poismg_mod
	124
[1320]	125	USE statistics, &
	126	ONLY: statistic_regions, sums_divnew_l, sums_divold_l, weight_pres, &
	127	weight_substep
	128
[1]	129	IMPLICIT NONE
	130
[1682]	131	INTEGER(iwp) :: i !<
	132	INTEGER(iwp) :: j !<
	133	INTEGER(iwp) :: k !<
[1]	134
[1682]	135	REAL(wp) :: ddt_3d !<
	136	REAL(wp) :: localsum !<
	137	REAL(wp) :: threadsum !<
	138	REAL(wp) :: d_weight_pres !<
[1]	139
[1682]	140	REAL(wp), DIMENSION(1:2) :: volume_flow_l !<
	141	REAL(wp), DIMENSION(1:2) :: volume_flow_offset !<
	142	REAL(wp), DIMENSION(1:nzt) :: w_l !<
	143	REAL(wp), DIMENSION(1:nzt) :: w_l_l !<
[1]	144
	145
	146	CALL cpu_log( log_point(8), 'pres', 'start' )
	147
[85]	148
[1342]	149	ddt_3d = 1.0_wp / dt_3d
	150	d_weight_pres = 1.0_wp / weight_pres(intermediate_timestep_count)
[85]	151
[1]	152	!
[707]	153	!-- Multigrid method expects array d to have one ghost layer.
	154	!--
[1575]	155	IF ( psolver(1:9) == 'multigrid' ) THEN
[667]	156
[1]	157	DEALLOCATE( d )
[667]	158	ALLOCATE( d(nzb:nzt+1,nys-1:nyn+1,nxl-1:nxr+1) )
[707]	159
	160	!
	161	!-- Since p is later used to hold the weighted average of the substeps, it
	162	!-- cannot be used in the iterative solver. Therefore, its initial value is
	163	!-- stored on p_loc, which is then iteratively advanced in every substep.
	164	IF ( intermediate_timestep_count == 1 ) THEN
	165	DO i = nxl-1, nxr+1
	166	DO j = nys-1, nyn+1
	167	DO k = nzb, nzt+1
	168	p_loc(k,j,i) = p(k,j,i)
	169	ENDDO
	170	ENDDO
	171	ENDDO
[667]	172	ENDIF
	173
[707]	174	ELSEIF ( psolver == 'sor' .AND. intermediate_timestep_count == 1 ) THEN
	175
	176	!
	177	!-- Since p is later used to hold the weighted average of the substeps, it
	178	!-- cannot be used in the iterative solver. Therefore, its initial value is
	179	!-- stored on p_loc, which is then iteratively advanced in every substep.
	180	p_loc = p
	181
[1]	182	ENDIF
	183
	184	!
[75]	185	!-- Conserve the volume flow at the outflow in case of non-cyclic lateral
	186	!-- boundary conditions
[106]	187	!-- WARNING: so far, this conservation does not work at the left/south
	188	!-- boundary if the topography at the inflow differs from that at the
	189	!-- outflow! For this case, volume_flow_area needs adjustment!
	190	!
	191	!-- Left/right
[709]	192	IF ( conserve_volume_flow .AND. ( outflow_l .OR. outflow_r ) ) THEN
[680]	193
[1342]	194	volume_flow(1) = 0.0_wp
	195	volume_flow_l(1) = 0.0_wp
[106]	196
	197	IF ( outflow_l ) THEN
	198	i = 0
	199	ELSEIF ( outflow_r ) THEN
	200	i = nx+1
	201	ENDIF
	202
	203	DO j = nys, nyn
	204	!
	205	!-- Sum up the volume flow through the south/north boundary
[709]	206	DO k = nzb_2d(j,i)+1, nzt
[667]	207	volume_flow_l(1) = volume_flow_l(1) + u(k,j,i) * dzw(k)
[106]	208	ENDDO
	209	ENDDO
	210
	211	#if defined( __parallel )
[680]	212	IF ( collective_wait ) CALL MPI_BARRIER( comm1dy, ierr )
[106]	213	CALL MPI_ALLREDUCE( volume_flow_l(1), volume_flow(1), 1, MPI_REAL, &
	214	MPI_SUM, comm1dy, ierr )
	215	#else
	216	volume_flow = volume_flow_l
	217	#endif
[709]	218	volume_flow_offset(1) = ( volume_flow_initial(1) - volume_flow(1) ) &
[106]	219	/ volume_flow_area(1)
	220
[667]	221	DO j = nysg, nyng
[709]	222	DO k = nzb_2d(j,i)+1, nzt
[106]	223	u(k,j,i) = u(k,j,i) + volume_flow_offset(1)
	224	ENDDO
	225	ENDDO
	226
	227	ENDIF
	228
	229	!
	230	!-- South/north
[709]	231	IF ( conserve_volume_flow .AND. ( outflow_n .OR. outflow_s ) ) THEN
[106]	232
[1342]	233	volume_flow(2) = 0.0_wp
	234	volume_flow_l(2) = 0.0_wp
[75]	235
[106]	236	IF ( outflow_s ) THEN
	237	j = 0
	238	ELSEIF ( outflow_n ) THEN
[75]	239	j = ny+1
[106]	240	ENDIF
	241
	242	DO i = nxl, nxr
[75]	243	!
[106]	244	!-- Sum up the volume flow through the south/north boundary
[709]	245	DO k = nzb_2d(j,i)+1, nzt
[667]	246	volume_flow_l(2) = volume_flow_l(2) + v(k,j,i) * dzw(k)
[75]	247	ENDDO
[106]	248	ENDDO
	249
[75]	250	#if defined( __parallel )
[680]	251	IF ( collective_wait ) CALL MPI_BARRIER( comm1dx, ierr )
[75]	252	CALL MPI_ALLREDUCE( volume_flow_l(2), volume_flow(2), 1, MPI_REAL, &
	253	MPI_SUM, comm1dx, ierr )
	254	#else
	255	volume_flow = volume_flow_l
	256	#endif
	257	volume_flow_offset(2) = ( volume_flow_initial(2) - volume_flow(2) ) &
[106]	258	/ volume_flow_area(2)
[75]	259
[667]	260	DO i = nxlg, nxrg
[709]	261	DO k = nzb_v_inner(j,i)+1, nzt
[106]	262	v(k,j,i) = v(k,j,i) + volume_flow_offset(2)
[75]	263	ENDDO
[106]	264	ENDDO
[75]	265
	266	ENDIF
	267
[76]	268	!
	269	!-- Remove mean vertical velocity
	270	IF ( ibc_p_b == 1 .AND. ibc_p_t == 1 ) THEN
[1342]	271	IF ( simulated_time > 0.0_wp ) THEN ! otherwise nzb_w_inner not yet known
	272	w_l = 0.0_wp; w_l_l = 0.0_wp
[76]	273	DO i = nxl, nxr
	274	DO j = nys, nyn
	275	DO k = nzb_w_inner(j,i)+1, nzt
	276	w_l_l(k) = w_l_l(k) + w(k,j,i)
	277	ENDDO
	278	ENDDO
	279	ENDDO
	280	#if defined( __parallel )
[622]	281	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
[709]	282	CALL MPI_ALLREDUCE( w_l_l(1), w_l(1), nzt, MPI_REAL, MPI_SUM, &
	283	comm2d, ierr )
[76]	284	#else
	285	w_l = w_l_l
	286	#endif
	287	DO k = 1, nzt
	288	w_l(k) = w_l(k) / ngp_2dh_outer(k,0)
	289	ENDDO
[667]	290	DO i = nxlg, nxrg
	291	DO j = nysg, nyng
[76]	292	DO k = nzb_w_inner(j,i)+1, nzt
	293	w(k,j,i) = w(k,j,i) - w_l(k)
	294	ENDDO
	295	ENDDO
	296	ENDDO
	297	ENDIF
	298	ENDIF
[75]	299
	300	!
[1]	301	!-- Compute the divergence of the provisional velocity field.
	302	CALL cpu_log( log_point_s(1), 'divergence', 'start' )
	303
[1575]	304	IF ( psolver(1:9) == 'multigrid' ) THEN
[1]	305	!$OMP PARALLEL DO SCHEDULE( STATIC )
	306	DO i = nxl-1, nxr+1
	307	DO j = nys-1, nyn+1
	308	DO k = nzb, nzt+1
[1342]	309	d(k,j,i) = 0.0_wp
[1]	310	ENDDO
	311	ENDDO
	312	ENDDO
	313	ELSE
	314	!$OMP PARALLEL DO SCHEDULE( STATIC )
[1111]	315	!$acc kernels present( d )
[1003]	316	DO i = nxl, nxr
	317	DO j = nys, nyn
	318	DO k = nzb+1, nzt
[1342]	319	d(k,j,i) = 0.0_wp
[1]	320	ENDDO
	321	ENDDO
	322	ENDDO
[1111]	323	!$acc end kernels
[1]	324	ENDIF
	325
[1342]	326	localsum = 0.0_wp
	327	threadsum = 0.0_wp
[1]	328
	329	#if defined( __ibm )
	330	!$OMP PARALLEL PRIVATE (i,j,k) FIRSTPRIVATE(threadsum) REDUCTION(+:localsum)
	331	!$OMP DO SCHEDULE( STATIC )
	332	DO i = nxl, nxr
	333	DO j = nys, nyn
	334	DO k = nzb_s_inner(j,i)+1, nzt
[85]	335	d(k,j,i) = ( ( u(k,j,i+1) - u(k,j,i) ) * ddx + &
	336	( v(k,j+1,i) - v(k,j,i) ) * ddy + &
[673]	337	( w(k,j,i) - w(k-1,j,i) ) * ddzw(k) ) * ddt_3d &
	338	* d_weight_pres
[1]	339	ENDDO
	340	!
	341	!-- Compute possible PE-sum of divergences for flow_statistics
	342	DO k = nzb_s_inner(j,i)+1, nzt
	343	threadsum = threadsum + ABS( d(k,j,i) )
	344	ENDDO
	345
	346	ENDDO
	347	ENDDO
	348
[707]	349	localsum = localsum + threadsum * dt_3d * &
	350	weight_pres(intermediate_timestep_count)
[693]	351
[1]	352	!$OMP END PARALLEL
	353	#else
	354
[1111]	355	!$OMP PARALLEL PRIVATE (i,j,k)
	356	!$OMP DO SCHEDULE( STATIC )
[1221]	357	!$acc kernels present( d, ddzw, rflags_s_inner, u, v, w )
	358	!$acc loop collapse( 3 )
[1111]	359	DO i = nxl, nxr
	360	DO j = nys, nyn
	361	DO k = 1, nzt
[1221]	362	d(k,j,i) = ( ( u(k,j,i+1) - u(k,j,i) ) * ddx + &
	363	( v(k,j+1,i) - v(k,j,i) ) * ddy + &
	364	( w(k,j,i) - w(k-1,j,i) ) * ddzw(k) ) * ddt_3d &
	365	* d_weight_pres * rflags_s_inner(k,j,i)
[1]	366	ENDDO
	367	ENDDO
[1111]	368	ENDDO
	369	!$acc end kernels
	370	!$OMP END PARALLEL
[1]	371
	372	!
	373	!-- Compute possible PE-sum of divergences for flow_statistics
	374	!$OMP PARALLEL PRIVATE (i,j,k) FIRSTPRIVATE(threadsum) REDUCTION(+:localsum)
	375	!$OMP DO SCHEDULE( STATIC )
[1221]	376	!$acc parallel loop collapse(3) present( d ) reduction(+:threadsum)
[1]	377	DO i = nxl, nxr
	378	DO j = nys, nyn
	379	DO k = nzb+1, nzt
	380	threadsum = threadsum + ABS( d(k,j,i) )
	381	ENDDO
	382	ENDDO
	383	ENDDO
[1257]	384	!$acc end parallel loop
[707]	385	localsum = localsum + threadsum * dt_3d * &
	386	weight_pres(intermediate_timestep_count)
[1]	387	!$OMP END PARALLEL
	388	#endif
	389
	390	!
	391	!-- For completeness, set the divergence sum of all statistic regions to those
	392	!-- of the total domain
	393	sums_divold_l(0:statistic_regions) = localsum
	394
	395	CALL cpu_log( log_point_s(1), 'divergence', 'stop' )
	396
	397	!
	398	!-- Compute the pressure perturbation solving the Poisson equation
[1575]	399	IF ( psolver == 'poisfft' ) THEN
[1]	400
	401	!
	402	!-- Solve Poisson equation via FFT and solution of tridiagonal matrices
[1575]	403	CALL poisfft( d )
[1212]	404
[1]	405	!
	406	!-- Store computed perturbation pressure and set boundary condition in
	407	!-- z-direction
	408	!$OMP PARALLEL DO
[1113]	409	!$acc kernels present( d, tend )
[1]	410	DO i = nxl, nxr
	411	DO j = nys, nyn
	412	DO k = nzb+1, nzt
	413	tend(k,j,i) = d(k,j,i)
	414	ENDDO
	415	ENDDO
	416	ENDDO
[1113]	417	!$acc end kernels
[1]	418
	419	!
	420	!-- Bottom boundary:
	421	!-- This condition is only required for internal output. The pressure
	422	!-- gradient (dp(nzb+1)-dp(nzb))/dz is not used anywhere else.
	423	IF ( ibc_p_b == 1 ) THEN
	424	!
	425	!-- Neumann (dp/dz = 0)
	426	!$OMP PARALLEL DO
[1113]	427	!$acc kernels present( nzb_s_inner, tend )
[667]	428	DO i = nxlg, nxrg
	429	DO j = nysg, nyng
[1]	430	tend(nzb_s_inner(j,i),j,i) = tend(nzb_s_inner(j,i)+1,j,i)
	431	ENDDO
	432	ENDDO
[1113]	433	!$acc end kernels
[1]	434
	435	ELSE
	436	!
	437	!-- Dirichlet
	438	!$OMP PARALLEL DO
[1113]	439	!$acc kernels present( tend )
[667]	440	DO i = nxlg, nxrg
	441	DO j = nysg, nyng
[1342]	442	tend(nzb_s_inner(j,i),j,i) = 0.0_wp
[1]	443	ENDDO
	444	ENDDO
[1113]	445	!$acc end kernels
[1]	446
	447	ENDIF
	448
	449	!
	450	!-- Top boundary
	451	IF ( ibc_p_t == 1 ) THEN
	452	!
	453	!-- Neumann
	454	!$OMP PARALLEL DO
[1113]	455	!$acc kernels present( tend )
[667]	456	DO i = nxlg, nxrg
	457	DO j = nysg, nyng
[1]	458	tend(nzt+1,j,i) = tend(nzt,j,i)
	459	ENDDO
	460	ENDDO
[1113]	461	!$acc end kernels
[1]	462
	463	ELSE
	464	!
	465	!-- Dirichlet
	466	!$OMP PARALLEL DO
[1113]	467	!$acc kernels present( tend )
[667]	468	DO i = nxlg, nxrg
	469	DO j = nysg, nyng
[1342]	470	tend(nzt+1,j,i) = 0.0_wp
[1]	471	ENDDO
	472	ENDDO
[1113]	473	!$acc end kernels
[1]	474
	475	ENDIF
	476
	477	!
	478	!-- Exchange boundaries for p
[1113]	479	IF ( numprocs == 1 ) THEN ! workaround for single-core GPU runs
	480	on_device = .TRUE. ! to be removed after complete porting
	481	ELSE ! of ghost point exchange
	482	!$acc update host( tend )
	483	ENDIF
[667]	484	CALL exchange_horiz( tend, nbgp )
[1113]	485	IF ( numprocs == 1 ) THEN ! workaround for single-core GPU runs
	486	on_device = .FALSE. ! to be removed after complete porting
	487	ELSE ! of ghost point exchange
	488	!$acc update device( tend )
	489	ENDIF
[1]	490
	491	ELSEIF ( psolver == 'sor' ) THEN
	492
	493	!
	494	!-- Solve Poisson equation for perturbation pressure using SOR-Red/Black
	495	!-- scheme
[707]	496	CALL sor( d, ddzu_pres, ddzw, p_loc )
	497	tend = p_loc
[1]	498
[1575]	499	ELSEIF ( psolver(1:9) == 'multigrid' ) THEN
[1]	500
	501	!
	502	!-- Solve Poisson equation for perturbation pressure using Multigrid scheme,
[667]	503	!-- array tend is used to store the residuals, logical exchange_mg is used
	504	!-- to discern data exchange in multigrid ( 1 ghostpoint ) and normal grid
	505	!-- ( nbgp ghost points ).
[778]	506
	507	!-- If the number of grid points of the gathered grid, which is collected
	508	!-- on PE0, is larger than the number of grid points of an PE, than array
	509	!-- tend will be enlarged.
	510	IF ( gathered_size > subdomain_size ) THEN
	511	DEALLOCATE( tend )
	512	ALLOCATE( tend(nzb:nzt_mg(mg_switch_to_pe0_level)+1,nys_mg( &
	513	mg_switch_to_pe0_level)-1:nyn_mg(mg_switch_to_pe0_level)+1,&
	514	nxl_mg(mg_switch_to_pe0_level)-1:nxr_mg( &
	515	mg_switch_to_pe0_level)+1) )
	516	ENDIF
	517
[1575]	518	IF ( psolver == 'multigrid' ) THEN
	519	CALL poismg( tend )
	520	ELSE
	521	CALL poismg_fast( tend )
	522	ENDIF
[707]	523
[778]	524	IF ( gathered_size > subdomain_size ) THEN
	525	DEALLOCATE( tend )
	526	ALLOCATE( tend(nzb:nzt+1,nysg:nyng,nxlg:nxrg) )
	527	ENDIF
	528
[1]	529	!
	530	!-- Restore perturbation pressure on tend because this array is used
	531	!-- further below to correct the velocity fields
[707]	532	DO i = nxl-1, nxr+1
	533	DO j = nys-1, nyn+1
	534	DO k = nzb, nzt+1
	535	tend(k,j,i) = p_loc(k,j,i)
	536	ENDDO
	537	ENDDO
	538	ENDDO
[667]	539
[1]	540	ENDIF
	541
	542	!
[707]	543	!-- Store perturbation pressure on array p, used for pressure data output.
	544	!-- Ghost layers are added in the output routines (except sor-method: see below)
	545	IF ( intermediate_timestep_count == 1 ) THEN
	546	!$OMP PARALLEL PRIVATE (i,j,k)
	547	!$OMP DO
[1113]	548	!$acc kernels present( p, tend, weight_substep )
[1257]	549	!$acc loop independent
[707]	550	DO i = nxl-1, nxr+1
[1257]	551	!$acc loop independent
[707]	552	DO j = nys-1, nyn+1
[1257]	553	!$acc loop independent
[707]	554	DO k = nzb, nzt+1
	555	p(k,j,i) = tend(k,j,i) * &
	556	weight_substep(intermediate_timestep_count)
[673]	557	ENDDO
[1]	558	ENDDO
[707]	559	ENDDO
[1113]	560	!$acc end kernels
[707]	561	!$OMP END PARALLEL
	562
	563	ELSE
	564	!$OMP PARALLEL PRIVATE (i,j,k)
	565	!$OMP DO
[1113]	566	!$acc kernels present( p, tend, weight_substep )
[1257]	567	!$acc loop independent
[707]	568	DO i = nxl-1, nxr+1
[1257]	569	!$acc loop independent
[707]	570	DO j = nys-1, nyn+1
[1257]	571	!$acc loop independent
[707]	572	DO k = nzb, nzt+1
	573	p(k,j,i) = p(k,j,i) + tend(k,j,i) * &
	574	weight_substep(intermediate_timestep_count)
[673]	575	ENDDO
	576	ENDDO
[707]	577	ENDDO
[1113]	578	!$acc end kernels
[707]	579	!$OMP END PARALLEL
	580
	581	ENDIF
[673]	582
[707]	583	!
	584	!-- SOR-method needs ghost layers for the next timestep
	585	IF ( psolver == 'sor' ) CALL exchange_horiz( p, nbgp )
[682]	586
[1]	587	!
	588	!-- Correction of the provisional velocities with the current perturbation
	589	!-- pressure just computed
[709]	590	IF ( conserve_volume_flow .AND. ( bc_lr_cyc .OR. bc_ns_cyc ) ) THEN
[1342]	591	volume_flow_l(1) = 0.0_wp
	592	volume_flow_l(2) = 0.0_wp
[1]	593	ENDIF
[707]	594
[1]	595	!$OMP PARALLEL PRIVATE (i,j,k)
	596	!$OMP DO
[1113]	597	!$acc kernels present( ddzu, nzb_u_inner, nzb_v_inner, nzb_w_inner, tend, u, v, w, weight_pres )
[1257]	598	!$acc loop independent
[673]	599	DO i = nxl, nxr
[1257]	600	!$acc loop independent
[1]	601	DO j = nys, nyn
[1257]	602	!$acc loop independent
[1113]	603	DO k = 1, nzt
	604	IF ( k > nzb_w_inner(j,i) ) THEN
	605	w(k,j,i) = w(k,j,i) - dt_3d * &
	606	( tend(k+1,j,i) - tend(k,j,i) ) * ddzu(k+1) * &
	607	weight_pres(intermediate_timestep_count)
	608	ENDIF
[1]	609	ENDDO
[1257]	610	!$acc loop independent
[1113]	611	DO k = 1, nzt
	612	IF ( k > nzb_u_inner(j,i) ) THEN
	613	u(k,j,i) = u(k,j,i) - dt_3d * &
	614	( tend(k,j,i) - tend(k,j,i-1) ) * ddx * &
	615	weight_pres(intermediate_timestep_count)
	616	ENDIF
[1]	617	ENDDO
[1257]	618	!$acc loop independent
[1113]	619	DO k = 1, nzt
	620	IF ( k > nzb_v_inner(j,i) ) THEN
	621	v(k,j,i) = v(k,j,i) - dt_3d * &
	622	( tend(k,j,i) - tend(k,j-1,i) ) * ddy * &
	623	weight_pres(intermediate_timestep_count)
	624	ENDIF
[673]	625	ENDDO
[1]	626
	627	ENDDO
	628	ENDDO
[1113]	629	!$acc end kernels
[1]	630	!$OMP END PARALLEL
[1113]	631
	632	!
	633	!-- Sum up the volume flow through the right and north boundary
	634	IF ( conserve_volume_flow .AND. bc_lr_cyc .AND. bc_ns_cyc .AND. &
	635	nxr == nx ) THEN
	636
	637	!$OMP PARALLEL PRIVATE (j,k)
	638	!$OMP DO
	639	DO j = nys, nyn
	640	!$OMP CRITICAL
	641	DO k = nzb_2d(j,nx) + 1, nzt
	642	volume_flow_l(1) = volume_flow_l(1) + u(k,j,nx) * dzw(k)
	643	ENDDO
	644	!$OMP END CRITICAL
	645	ENDDO
	646	!$OMP END PARALLEL
	647
	648	ENDIF
	649
	650	IF ( conserve_volume_flow .AND. bc_ns_cyc .AND. bc_lr_cyc .AND. &
	651	nyn == ny ) THEN
	652
	653	!$OMP PARALLEL PRIVATE (i,k)
	654	!$OMP DO
	655	DO i = nxl, nxr
	656	!$OMP CRITICAL
	657	DO k = nzb_2d(ny,i) + 1, nzt
	658	volume_flow_l(2) = volume_flow_l(2) + v(k,ny,i) * dzw(k)
	659	ENDDO
	660	!$OMP END CRITICAL
	661	ENDDO
	662	!$OMP END PARALLEL
	663
	664	ENDIF
[673]	665
[1]	666	!
	667	!-- Conserve the volume flow
[707]	668	IF ( conserve_volume_flow .AND. ( bc_lr_cyc .AND. bc_ns_cyc ) ) THEN
[1]	669
	670	#if defined( __parallel )
[622]	671	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
[1]	672	CALL MPI_ALLREDUCE( volume_flow_l(1), volume_flow(1), 2, MPI_REAL, &
	673	MPI_SUM, comm2d, ierr )
	674	#else
	675	volume_flow = volume_flow_l
	676	#endif
	677
	678	volume_flow_offset = ( volume_flow_initial - volume_flow ) / &
	679	volume_flow_area
	680
	681	!$OMP PARALLEL PRIVATE (i,j,k)
	682	!$OMP DO
	683	DO i = nxl, nxr
	684	DO j = nys, nyn
[667]	685	DO k = nzb_u_inner(j,i) + 1, nzt
	686	u(k,j,i) = u(k,j,i) + volume_flow_offset(1)
[719]	687	ENDDO
	688	DO k = nzb_v_inner(j,i) + 1, nzt
[667]	689	v(k,j,i) = v(k,j,i) + volume_flow_offset(2)
	690	ENDDO
[1]	691	ENDDO
	692	ENDDO
[667]	693
[1]	694	!$OMP END PARALLEL
	695
	696	ENDIF
	697
	698	!
	699	!-- Exchange of boundaries for the velocities
[1113]	700	IF ( numprocs == 1 ) THEN ! workaround for single-core GPU runs
	701	on_device = .TRUE. ! to be removed after complete porting
	702	ELSE ! of ghost point exchange
	703	!$acc update host( u, v, w )
	704	ENDIF
[667]	705	CALL exchange_horiz( u, nbgp )
	706	CALL exchange_horiz( v, nbgp )
	707	CALL exchange_horiz( w, nbgp )
[1113]	708	IF ( numprocs == 1 ) THEN ! workaround for single-core GPU runs
	709	on_device = .FALSE. ! to be removed after complete porting
	710	ELSE ! of ghost point exchange
	711	!$acc update device( u, v, w )
	712	ENDIF
[1]	713
	714	!
	715	!-- Compute the divergence of the corrected velocity field,
	716	!-- a possible PE-sum is computed in flow_statistics
	717	CALL cpu_log( log_point_s(1), 'divergence', 'start' )
[1342]	718	sums_divnew_l = 0.0_wp
[1]	719
	720	!
	721	!-- d must be reset to zero because it can contain nonzero values below the
	722	!-- topography
[1342]	723	IF ( topography /= 'flat' ) d = 0.0_wp
[1]	724
[1342]	725	localsum = 0.0_wp
	726	threadsum = 0.0_wp
[1]	727
	728	!$OMP PARALLEL PRIVATE (i,j,k) FIRSTPRIVATE(threadsum) REDUCTION(+:localsum)
	729	!$OMP DO SCHEDULE( STATIC )
	730	#if defined( __ibm )
	731	DO i = nxl, nxr
	732	DO j = nys, nyn
	733	DO k = nzb_s_inner(j,i)+1, nzt
	734	d(k,j,i) = ( u(k,j,i+1) - u(k,j,i) ) * ddx + &
	735	( v(k,j+1,i) - v(k,j,i) ) * ddy + &
	736	( w(k,j,i) - w(k-1,j,i) ) * ddzw(k)
	737	ENDDO
	738	DO k = nzb+1, nzt
	739	threadsum = threadsum + ABS( d(k,j,i) )
	740	ENDDO
	741	ENDDO
	742	ENDDO
	743	#else
[1221]	744	!$acc kernels present( d, ddzw, rflags_s_inner, u, v, w )
	745	!$acc loop collapse( 3 )
[1]	746	DO i = nxl, nxr
	747	DO j = nys, nyn
[1113]	748	DO k = 1, nzt
[1221]	749	d(k,j,i) = ( ( u(k,j,i+1) - u(k,j,i) ) * ddx + &
	750	( v(k,j+1,i) - v(k,j,i) ) * ddy + &
	751	( w(k,j,i) - w(k-1,j,i) ) * ddzw(k) &
	752	) * rflags_s_inner(k,j,i)
[1113]	753	ENDDO
	754	ENDDO
	755	ENDDO
	756	!$acc end kernels
	757	!
	758	!-- Compute possible PE-sum of divergences for flow_statistics
[1221]	759	!$acc parallel loop collapse(3) present( d ) reduction(+:threadsum)
[1113]	760	DO i = nxl, nxr
	761	DO j = nys, nyn
[1221]	762	DO k = nzb+1, nzt
[1]	763	threadsum = threadsum + ABS( d(k,j,i) )
	764	ENDDO
	765	ENDDO
	766	ENDDO
[1257]	767	!$acc end parallel loop
[1]	768	#endif
[667]	769
[1]	770	localsum = localsum + threadsum
	771	!$OMP END PARALLEL
	772
	773	!
	774	!-- For completeness, set the divergence sum of all statistic regions to those
	775	!-- of the total domain
	776	sums_divnew_l(0:statistic_regions) = localsum
	777
	778	CALL cpu_log( log_point_s(1), 'divergence', 'stop' )
	779
	780	CALL cpu_log( log_point(8), 'pres', 'stop' )
	781
	782
	783	END SUBROUTINE pres

Note: See TracBrowser for help on using the repository browser.

Download in other formats:

| Impressum | ©Leibniz Universität Hannover |